Discrete Optimum Design of Planar Steel Curved Roof and Pitched Roof Portal Frames Using Metaheuristic Algorithms

Portal frames are single-story frame buildings including columns and rafters, and their rafters can be either curved or pitched. These are used widely in the construction of industrial buildings, warehouses, gyms, fire stations, agricultural buildings, hangars, etc. The construction cost of these frames considerably depends on their weight. In the present research, the discrete optimum design of two types of portal frames including planar steel Curved Roof Frame (CRF) and Pitched Roof Frame (PRF) with tapered I-section members are presented. The optimal design aims to minimize the weight of these frame structures while satisfying some design constraints based on the requirements of ANSI/AISC 360-16 and ASCE 7-10. Four population-based metaheuristic optimization algorithms are applied to the optimal design of these frames. These algorithms consist of Teaching-Learning-Based Optimization (TLBO), Enhanced Colliding Bodies Optimization (ECBO), Shuffled Shepherd Optimization Algorithm (SSOA), and Water Strider Algorithm (WSA). Two main objectives are followed in this paper. The first one deals with comparing the optimized weight of the CRF and PRF structures with the same dimensions for height and span in two different span lengths (16.0 m and 32.0 m), and the second one is related to comparing the performance of the considered metaheuristics in the optimum design of these portal frames. The obtained results reveal that CRF is more economical than PRF in the fair comparison. Moreover, comparing the results acquired by SSOA with those of other considered metaheuristics reveals that SSOA has better performance for the optimal design of these portal frames.

Discrete Optimum Design of Truss Structures by an Improved Firefly Algorithm

This paper presents an improved firefly algorithm (FA) for fast optimization of truss structures with discrete variables. The enhanced accelerated firefly algorithm (AFA) is a simple, but very effective modification of FA. In order to investigate the performance and robustness of the proposed algorithm, some benchmark (structural optimization) problems are solved and the results are compared with FA and other algorithms. The results show that in some test cases, AFA not only finds lighter structures compared to other algorithms, but also converges faster. In the rest test cases, the optimal solutions are found with very less computational effort. The study also shows that the proposed AFA remarkably improves stability of the firefly algorithm in discrete design of truss structures.

The Research of Discrete Optimum Design of Composite Laminate Materials Using Lamination Parameters

Since lamination parameters were found to represent the stiffness properties and the orientation of each ply, two major stages are defined to do the optimization. In the first stage which is called continuous optimization, the required lamination parameters will be obtained by VICONOPT but this has not been completed. This paper focuses on the second stage which is called discrete optimization. It consists of three programs and the major program is called OPBDP. Random values of required lamination parameters for 4, 8 and 16 plies lay-ups are chosen and assumed, then the results are computed and compared. What is more, a method called double up the plies is introduced and the results show that it could save a lot of solution time but is still only suitable for a small number of plies.